We distinguish between recognition of familiar faces, which is not marked by task achievement, and therefore does not admit of recognitional failure, and cases of facial recognition which, in normal circumstances, are marked by task achievement insofar as there is a presumption that one may not know who or what one is currently perceiving. We expect recognitional failure if one has not seen an old friend for twenty years, and now she is before one, for she has probably changed greatly since the previous encounter, or we presume that one may not know the facial expression I immediately recognize now if the original encounter of it was a quick look, and so on. Unfortunately, the philosophers who conceive of recognition as consisting in a hidden process of coinciding one’s impression with a stored representation easily gloss over these subtle differences, and build into their ancient philosophy the mistaken claim that whenever one perceives a familiar face, one must recognize it.

The philosopher who asks me whether I recognize my wife’s familiar smile in the photo would receive a positive reply. Now, I might add to this some description or other, such as: ‘Her face radiates joy’, or ‘What a face – it says something!’. It is quite natural to do so. Suppose, however, that the philosopher decides to press me further about the meaning in her face by asking a question such as, ‘What does her face say?’. Is this clear? What should I say in response? In normal circumstances, the words in the description of a familiar face are not meant as a preliminary to specifying what meaning it has. When we find a human face noteworthy, be it a familiar one or not, we are interested in what it expresses, we let it impress itself upon us, and we convey in language how impressed we are by, how arresting we find, the object of our attention. Naturally, we may say in this context: ‘That face has a friendly meaning’ or, ‘Now that is sadness’, and so on. If the philosopher now asks ‘What does the face say?’, we find that no provision is normally made for such a question. In fact, we are merely ‘giving ourselves up to the features’, Wittgenstein cautions, letting them impress us. The philosopher makes the mistake of always looking at the language of human facial expression according to the same preconceived schema: as meaning the outward manifestation of a hidden process.

Consider music as an analogy to this case. When we let ourselves be impressed by music, we may remark ‘This music says something‘. We are strongly inclined in philosophy to think that one should be able to specify what it says. ‘Music is love in search of a word’, Sidney Lanier once said; but clearly, this is not always the case. For, I do not always mean the words to preface a description of my imaginative involvement. Rather, they function merely to convey my level of interest: the music really did impress me. When words lead off, music begins. Now, we can imagine cases where the use of ‘such-and-such made a quite particular impression’ is used as an introduction to an interpretation. We may say ‘It’s extremely impressive: the music drives on relentlessly, the horns sound off to increase the moment, and the strings bring the tension to a definitive close’. When words lead off, the music is in the language. But again, the transition made here from an expression of interest in the music to an interpretation is conducted in the language we speak, and the move does not involve transition to any mental process of understanding or recognition that supposedly underlies the experience, and that accompanies it on every occasion. If I ask for your interpretation in this case, it is because your level of interest in the music is noteworthy. It merits attention, a second glance, as it were, and our exchange is carried on within our language; and does not involve a transition to a mental process. My interpretation remains within language.

The expression on a familiar face is simply there, alive in the features. It is not concentrated or localized in the result of any process or mental event; it is not seen through. The expression is dispersed in the features of the face and embeds there like diffuse light (Wittgenstein, Remarks on the Philosophy of Psychology, Volume II; 1980). One may say that the face is familiar. But the familiarity, once more, is not a matter of a comparison made, for example, with a memory-image. It is not as though the image of a face is less familiar than a face itself. And if an interpretation of a familiar face calls to be explained, then the explanation of it is continuous with the language in which it is given, and does not betoken the presence of a process of interpretation or recognition. We move in language.